article

Site hosted by Angelfire.com: Build your free website today!

Child Abuse & Neglect
Volume 26, Issue 9, September 2002, Pages 939-953

A screening instrument for adults to assess past and current experiences of abuse

C. G. Kooiman, a, A. W. Ouwehanda and M. M. ter Kuileb

a Department of Psychiatry, B1-P, Leiden University Medical Center, Post Box 9600, 2300 RC, Leiden, The Netherlands
b Department of Gynecology, Outpatient Clinic of Psychosomatic Gynecology and Sexology, Leiden University Medical Center, Leiden, The Netherlands

Received 1 September 2000; revised 27 November 2001; accepted 9 December 2001. Available online 20 July 2002.

Abstract
Objective: We reconstructed and validated a simple questionnaire to be completed by adult respondents for the assessment of sexual and physical abuse during childhood and later life, the Sexual and Physical Abuse Questionnaire (SPAQ).

Method: The criterion validity of the questionnaire was investigated in a population of psychiatric outpatients (n=134) using the Structured Trauma Interview [Am. J. Psychiatr. 156 (1999) 379] as gold standard for the assessment of sexual and physical abuse.

Results: All questionnaires were returned fully completed. The measures of agreement and the predictive measures of the questionnaire were satisfactory, in particular with respect to experiences of sexual abuse. Positive answering of the questionnaire increased the odds for sexual abuse by a factor of 12¯17.5, and negative answering of the questionnaire reduced the odds by a third. The odds for physical abuse were increased by a factor of 8 with positive answering of the questionnaire, and reduced by a third with negative answering.

Conclusion: The questionnaire may be a useful screening instrument in research and in clinical practice to assess sexual abuse during childhood and later years. As a screening instrument for physical abuse it is less satisfactory.

Introduction
In order to assess a history of sexual and physical abuse, both structured interviews (Draijer and Ensink; Fink, Bernstein, Handelsman, Foote, & Lovejoy, 1995; Gallagher, Flye, Hurt, Stone, & Hull, 1992; Herman, Perry, & Van der Kolk, 1989; Römkens, 1992; Zanarini, Gunderson, Marino, Schwartz, & Frankenburg, 1989) and questionnaires ( Arntz, Dietzel, & Dreessen, 1999; Bernstein et al., 1994; Bryer, Nelson, Miller, & Kroll, 1987; Lange, Kooiman, Huberts, & Van Oostendorp, 1995; Nijenhuis, Spinhoven, Van Dyck, Van der Hart, & Vanderlinden, 1998) have been developed. Although there is no "gold standard" for establishing sexual and physical abuse, it is assumed that an interview is a more reliable and more valid method for determining the prevalence of sexual and physical abuse than a self-report questionnaire ( Wyatt & Peters, 1986). Structured interviews are, however, time-consuming, even when findings are negative. In large populations it would therefore be useful if a self-report questionnaire could be used as a screening instrument. However, most questionnaires are too detailed because they are meant as a substitute for a detailed interview, and others are limited to one specific aspect, usually sexual abuse, or to one particular period of the respondent's life, usually childhood.

Leserman, Drossman, and Li (1995) developed a practical screening instrument to be completed by adults for the assessment of sexual and physical abuse. The criterion validity of their screening instrument and the test¯retest reliability could be classified as satisfactory. The formulation of the questions concerning sexual abuse was determined by a broad definition of sexual abuse including all kinds of noncontact abuse such as unwanted exposure to someone's sexual body parts. However, for research we preferred a narrower definition limited to sexual abuse in which actual physical contact took place, because we wanted to focus on those events that could be classified as abusive without too much discussion about their content. Similarly, we considered the items on physical abuse to be too broadly defined, including for example having been incidentally hit or kicked without any physical harm. Finally, we preferred a more detailed age categorization than the one used by Leserman et al. (1995).

We therefore decided to modify the questionnaire developed by Leserman et al. (1995) in accordance with definitions ( Draijer and Walling) of sexual abuse limited to abuse with actual physical contact, and physical abuse limited to abuse resulting in physical harm such as bruises. Consequently, only those questions in the screening questionnaire developed by Leserman et al. (1995) relating to actual sexual contact were retained, and the questions about physical abuse were replaced with one question about physical violence purposefully inflicted on the respondent that resulted in physical harm. These are the core questions of the questionnaire. In order to meet the wishes of clinicians who prefer a screening instrument that also asks for less specific forms of abuse, we added two general questions on experiences that were in some way perceived as sexually abusive or violent. Also for clinical reasons three more questions were added which ask whether the subject had been able to discuss the experiences, if any, with other people and whether he or she presently has a need to discuss them. Finally, any experiences of sexual or physical abuse were subdivided according to Herman et al. (1989) into four age groups: early childhood (age <6 years), latency (6 yearsage<12 years), puberty (12 yearsage<16 years) and adulthood (age16 years).

This study investigates the validity of the questionnaire, the Sexual and Physical Abuse Questionnaire (SPAQ), by comparing the answers to the core questions of the SPAQ to the corresponding answers given during sexual and physical abuse.

Material and methods
Patient selection and procedures
Over a 9-month period, all consecutive new patients of the psychiatric outpatients' clinic of our hospital were invited to take part in the study. Patients under 16 years of age or above 64 years of age, those with an organic mental disorder, mental retardation, psychosis, sensory disturbances, or patients who did not have sufficient command of the Dutch language were excluded from the study.

The diagnostic phase of treatment consisted of two interviews separated by a 2-week period. During this 2-week gap an additional interview was carried out focusing on the occurrence of sexual and physical abuse. This interview was carried out after the patients had completed the SPAQ at home. The interviewers did not know the answers to the questionnaire.

One male and two female trained interviewers carried out the interviews. Patients were randomly assigned to one of the male or female interviewers, regardless of gender. The data obtained by interview and questionnaire were available for the treating psychiatrist to be used for the clinical evaluation unless the patient withheld his consent (which none of the patients did). These procedures are in conformity with common clinical practice and thereby assure the clinical validity of the results (Hogan & Nicholson, 1988).

Patients gave their informed consent after full verbal and written information about the nature and procedures of the study. The study was approved by the hospital's Committee of Medical Ethics.

Instruments
The treating psychiatrist noted the sociodemographic details (Jonsson and UNESCO) and the psychiatric diagnosis ( American Psychiatric Association, 1994) on a standardized form.

The Structured Trauma Interview (STI) (Draijer & Langeland, 1999) is a structured interview about sexual and physical abuse during childhood and later years. Sexual contact is defined in the interview as any actual physical sexual contact. Sexual abuse is defined as any sexual contact that occurred against the subject's will, or without the person involved feeling able to refuse the sexual contact. Sexual abuse by playmates during childhood is not included because of the ambivalence concerning the sexual contact that the respondent may have experienced. Physical abuse is defined as physical violence purposefully inflicted on the respondent as a result of which the respondent suffered physical injury. The term "physical injury" covers all form of injury ranging from bruises to broken bones and unconsciousness. The subject of physical abuse during childhood is limited in the interview to abuse by the parents or caregivers only. Regular meetings were held between the interviewers to ensure maximum consistency in the way the interviews were carried out and scored.

The Sexual and Physical Abuse Questionnaire has already been described in the introduction and is represented in the Appendix A. For this study a Dutch version of the SPAQ is used. Only the results of the core questions concerning sexual (#1, #2, and #3) and physical abuse (#6) are used for the analysis of the criterion validity of the SPAQ with the STI.

Analysis of the data
The chi-square test is used to test for differences between groups. Cohen's weighted kappa is used as a conventional measure of agreement between questionnaire and interview. We calculated kappa for all age groups as well as for childhood abuse only. The kappa statistic treats all disagreements equally, for example `no sexual abuse on the SPAQ versus sexual abuse before the age of 6 years on the STI' and `sexual abuse at age 12¯16 years on the SPAQ versus sexual abuse at age 6¯12 years on the STI.' Therefore, we used the weighted kappa that also takes account of the degree of disagreement. Kappa values are interpreted as poor (.00¯.20), fair (.21¯.40), moderate (.41¯.60), good (.61¯.80) and excellent (.81¯1.00) (Altman, 1997). Statistical calculations are done with SPSS 8.0.0.

When the questionnaire is being viewed as a screening instrument and the interview as the gold standard, the sensitivity (sens), specificity (spec), positive predictive value (PVpos), and the negative predictive value (PVneg) of the SPAQ as a screening instrument are of importance. However, also the "likelihood ratios" for both positive and negative tests are presented. The likelihood ratio for a positive test is expressed as LR+=sens/(1-spec) and that of a negative test as LR=(1-sens)/spec. The likelihood ratios can be used to calculate the probability of sexual and physical abuse on the basis of the results of the questionnaire if the prevalence is known (by means of the interview). The chance of a random patient having a history of (sexual or physical) abuse can be expressed in terms of odds. The odds for abuse is the chance that abuse occurred divided by the chance that abuse did not occur. Assuming that the data from the interview in theory should reflect the "true" prevalence, the calculated odds for the interview are known as the "a priori odds." If the a priori odds are multiplied by the LR+ or LR- values, then one obtains the "a posteriori odds" given a positive or negative response to the questionnaire. Using the latter, the "a posteriori probability" of abuse can be calculated (odds/1+odds). The likelihood ratios give a better indication of the value of a test or questionnaire than the PVpos, PVneg or Cohen's kappa, since unlike these values, the likelihood ratios are independent of the prevalence of abuse and can therefore be used in any population (Altman, 1997). Likelihood ratios are calculated for childhood (sexual or physical) abuse, that is, reported abuse before the respondent reached the age of 16 years, and for lifetime abuse, that is, reported abuse at any age period.

Results
During the study period 171 patients were eligible for the study. Thirty-seven (22%) patients refused to participate. There was no difference between the participants and the 37 refusers in terms of age, "partnership," highest completed education, profession or psychiatric diagnosis. There were, however, more women in the group who refused to participate in the study (78% vs. 57%; 2=5.741, p<.05).

The study population consisted of 76 women and 58 men with an mean age of 39.4 (SD 11.8) were single; 21% only had primary school education, while 13% had a university education; 14% did not have a paid job, 35% were blue collar, and 51% were white collar workers; 34% of the patients had a somatoform disorder, 29% had an anxiety or mood disorder, 25% had another mental disorder, and 12% of the patients did not have a mental disorder.

Findings with the Sexual and Physical Abuse Questionnaire
All questionnaires were completely answered. When asked, the patients considered the questionnaire easy to read and none of the patients had difficulty filling in the SPAQ. Some patients used the open spaces in the questionnaire to add explanatory comments.

Sexual abuse (SPAQ)
Thirty women and seven men, in total 28% of the 134 patients, gave a positive answer to at least one of the four questions about sexual abuse. Four women and three men only gave a positive answer to the general question about sexual abuse. For three of these women it concerned incidents that occurred in the age group of 16 years or older, and for the three men the events occurred during puberty and involved sexually-charged situations which were perceived as unpleasant (exhibitionism) or threatening (sexual intimidation). One woman reported an extremely sexually threatening experience when aged less than 6 years old.

Twenty-six women and four men, in total 22% of all the patients, answered at least one of the three core questions about sexual abuse positively (Table 1). Women reported sexual abuse experiences more frequently than men did (p<.01). Twenty-three (17%) of the cases involved sexual abuse during childhood (< 16 year). Eleven patients having a history of sexual abuse had never discussed it with anyone before. Only one of these patients expressed the desire to discuss the abuse with one of the healthcare workers, and two of them never wanted to talk about it with anyone.

Physical abuse (SPAQ)
Twenty-three women and 21 men, in total 33% of the 134 patients, gave a positive answer to at least one of the two questions about physical abuse. Four women and four men only gave a positive answer to the general question. For both the men and the women this included experiences ranging from bullying and being hit without injury to experiences of severe threats of violence. These experiences were reported for all age categories.

Nineteen women and 17 men, representing 27% of all patients, gave a positive answer to the core question about physical abuse (Table 2). Women reported physical abuse as frequently as men. In 27 cases (20%) the experience of physical abuse occurred during childhood. Eight patients with a history of physical abuse had never discussed it with anyone before, two patients indicated that they did not wish to discuss it.

Findings using the STI
Sexual abuse (STI)
During the interview 34 (25%) of the 134 patients reported ever having had an experience of sexual abuse. For 27 patients (20%) this involved sexual abuse during childhood. In six of these cases it involved abuse by one of the parents (the father in all cases but one). Only two cases involved abuse by a stranger. Eight patients suffered the abuse for more than a year. Eight of the 27 patients with sexual abuse in childhood reported that at the time they did not experience, or barely experienced the sexual abuse as disturbing. Eight of these 27 patients were also victims of sexual violence in their adulthood.

Physical abuse (STI)
Forty-two patients (31%) reported ever having been the victim of a form of physical abuse. Twenty-four (18%) of the 134 patients experienced physical abuse during childhood. Two cases of child-abuse resulted in bone fractures, two cases involved dislocations and three involved burns. In 21 of the patients the physical abuse during childhood took place over more than 1 year. Fifteen of the 24 patients who experienced physical abuse during childhood were also victims of physical abuse later on, from 16 years of age onwards.

Degree of agreement between SPAQ and STI
The SPAQ revealed slightly lower prevalence values of sexual and physical abuse than the STI. For sexual abuse these were 22% (95% CI: 15.0¯29.0%) and 25% (95% CI: 17.7¯32.3%) in the SPAQ and STI, respectively, and for physical abuse the prevalence values were 27% (95% CI: 19.5¯34.5%) and 31% (95% CI: 23.2¯38.8%), respectively.
Discussion
The SPAQ is constructed as a self-report questionnaire to assess the prevalence of sexual as well as physical abuse during childhood and later years using narrow definitions of abuse. Sexual abuse being restricted to sexual abuse with actual physical contact and physical abuse being restricted to intentional violence resulting in some kind of physical injury as for example bruises.

With regard to sexual abuse the SPAQ has good levels of agreement with a structured interview in terms of Cohen's kappa. The predictive measures for sexual abuse as specificity, PVpos, PVneg, and the likelihood ratios can be qualified as good. Positive answering of one of the core questions on sexual abuse increased the odds for sexual abuse by a factor 12 (lifetime abuse) to 17.5 (childhood abuse), and negative answering of one of these questions reduced the odds by a third for lifetime as well as childhood sexual abuse. With regard to physical abuse the SPAQ has moderate (lifetime) to good (childhood) levels of agreement in terms of Cohen's kappa. The predictive measures of the SPAQ for physical abuse as specificity, PVneg, and the likelihood ratios can be qualified as good with the exception of PVpos for childhood physical abuse. The odds for lifetime as well as childhood physical abuse were increased with a factor 8 with positive answering of the core question, and reduced by a third with negative answering. In accordance with previous findings (Wyatt & Peters, 1986) the interview generally showed slightly higher prevalence rates for sexual and physical abuse than the questionnaire. Accordingly, the sensitivity values of the SPAQ were generally lower than the other predictive measures. In other words, people without a history of sexual or physical abuse generally don't score these on the questionnaire. However, there are some patients who do not report sexual or physical abuse on the questionnaire, but who do report it during the interview. On the basis of these results we conclude that the SPAQ may be a useful screening instrument for sexual abuse. It seems somewhat less suitable for use as a screening instrument for physical abuse.

Our study is one of very few studies on the validity of instruments to assess sexual and physical abuse. To our knowledge only two other studies investigated the criterion validity of a questionnaire on abuse. In both studies (Bernstein and Leserman) broader definitions were applied for sexual abuse, including abuse without physical contact, and for physical abuse, including abuse without physical injury. In a population of mainly male drug- or alcohol-dependent patients, Bernstein et al. (1994) found a high correlation (r=.65) between dimensional measures for childhood sexual abuse reported on their questionnaire and a standardized interview. The correlation (r=.38) of the measures for childhood physical abuse, although still significant, was considerably lower. Unfortunately, the raw data and other measures of convergence are not presented in their publication. However, one can infer from their data that sexual abuse was not very frequent in their population and this may have accentuated the high correlation found. Leserman et al. (1995) validated their questionnaire in a population of female gastro-enterology patients. They found a higher percentage of patients with sexual abuse (41% vs. 28%) and a higher percentage of patients with a history of physical abuse (47% vs. 33%) than we did. These differences in prevalence are probably due to a difference in the populations studied with a higher risk for sexual abuse in a sample of only females. The differences may be further explained by the stricter definitions of sexual and physical abuse applied in the present study. Unlike Leserman et al. (1995) we did not focus on forms of sexual abuse without any physical contact or physical abuse without any physical harm. It is noteworthy that both studies nevertheless showed the same sensitivity and specificity values, except for the specificity regarding physical abuse that was higher for the SPAQ.

More women than men refused to take part in the study. The clinicians treating the patients felt that in particular women with a traumatic history shrunk away from the study. This selection bias probably decreased the prevalence values found in our study. However, we consider it unlikely that the selection bias will have negatively influenced the degree of agreement between the SPAQ and the STI, although one may also speculate that persons, who refuse to participate in a study like this, are possibly more likely to give inconsistent reports. In the same line of reasoning the test¯retest reliability of the SPAQ is of interest. This, however, has to be established by further research.

We used a structured interview as the gold standard to assess sexual and physical abuse. However, although interviews are supposed to be more reliable than self-report questionnaires (Wyatt & Peters, 1986), it should be noticed that also interviews do not perfectly match reality. Some respondents may have suffered abuse without any conscious memory of it ( Williams, 1995), while other respondents wrongly assume to have suffered experiences of sexual and physical abuse ( Loftus, 1997). One may consider the use of independent sources of information such as hospital or police records as alternatives for the interview, but that does not seem to be a solution as sexual and physical abuse often are not discussed with others. In our study 11 of 30 (37%) patients never discussed before with another person their sexual abuse experiences and 8 of 37 (22%) patients never discussed with someone their experiences with physical abuse.

Discrepancies between the questionnaire and the interview are partly due to differences in emphasis in the two instruments. Particularly with regard to physical abuse in childhood there is a discrepancy between both instruments, because the interview is restricted to physical abuse by the parents only. Accordingly, in particular the PVpos for physical abuse during childhood, which of course could also be inflicted by people other than the parents, is suppressed as a result. Discrepancies may further be partly explained by the nature of the problem. Men in particular seem to be ashamed to indicate sexual abuse on the questionnaire, but nevertheless report it during the interview.

The SPAQ is expressly designed to be a screening instrument to assess the prevalence of sexual and physical abuse related to the age-period in both men and women. As a screening instrument it is not meant to replace a detailed interview. When used alone it may result in relatively high rates of false negative classifications. Taking this into consideration, the SPAQ appears to be a simple-to-use screening instrument with established likelihood ratios and measures of agreement. As such it can be used in scientific research as well as in clinical practice.

References
Altman, 1997. Altman, D. G. (1997). Practical statistics for medical research. London: Chapman & Hall.

American Psychiatric Association, 1994. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association.

Arntz et al., 1999. A. Arntz, R. Dietzel and L. Dreessen , Assumptions in borderline personality disorder: Specificity, stability and relationship with etiological factors. Behavior Research and Therapy 37 (1999), pp. 545¯557.

Bernstein et al., 1994. D.P. Bernstein, L. Fink, L. Handelsman, J. Foote, M. Lovejoy, K. Wenzel, E. Sapareto and J. Ruggiero , Initial reliability and validity of a new retrospective measure of child abuse and neglect. American Journal of Psychiatry 151 (1994), pp. 1132¯1136.

Bryer et al., 1987. J.B. Bryer, B.A. Nelson, J.B. Miller and P. Kroll , Childhood sexual and physical abuse as factors in adult psychiatric illness. American Journal of Psychiatry 144 (1987), pp. 1426¯1430.

Draijer, 1990. Draijer, N. (1990). Seksuele traumatisering in de jeugd. Gevolgen op lange termijn van seksueel misbruik van meisjes door verwanten. (Sexual traumatization during youth. Long term consequences of sexual abuse of young girls by relatives). Amsterdam: Uitgeverij SUA.

Draijer & Langeland, 1999. N. Draijer and W. Langeland , Childhood trauma and perceived parental dysfunction in the etiology of dissociative symptoms in psychiatric inpatients. American Journal of Psychiatry 156 (1999), pp. 379¯385.

Ensink, 1992. Ensink, B. J. (1992). Confusing realities. A study on child sexual abuse and psychiatric symptoms. Amsterdam: VU University Press.

Fink et al., 1995. L.A. Fink, D. Bernstein, L. Handelsman, J. Foote and M. Lovejoy , Initial reliability and validity of the childhood trauma interview: A new multidimensional measure of childhood interpersonal trauma. American Journal of Psychiatry 152 (1995), pp. 1329¯1335.

Gallagher et al., 1992. R.E. Gallagher, B.L. Flye, S.W. Hurt, M.H. Stone and J.W. Hull , Retrospective assessment of traumatic experiences (RATE). Journal of Personality Disorders 6 (1992), pp. 99¯108.

Herman et al., 1989. J.L. Herman, J.C. Perry and B. Van der Kolk , Childhood trauma in borderline personality disorder. American Journal of Psychiatry 146 (1989), pp. 490¯495.

Hogan & Nicholson, 1988. R. Hogan and R.A. Nicholson , The meaning of personality test scores. American Psychologist 43 (1988), pp. 621¯626.

Jonsson, 1987. J.O. Jonsson , Class origin, cultural origin, and educational attainment: The case of Sweden. European Sociological Review 3 (1987), pp. 229¯242.

Lange et al., 1995. A. Lange, K. Kooiman, L. Huberts and E. Van Oostendorp , Childhood unwanted sexual events and degree of psychopathology of psychiatric patients: Research data with a new anamnestic questionnaire (the CHUSE). Acta Psychiatrica Scandinavica 92 (1995), pp. 441¯446.

Leserman et al., 1995. J. Leserman, D.A. Drossman and Z. Li , The reliability and validity of a sexual and physical abuse history questionnaire in female patients with gastrointestinal disorders. Behavioral Medicine 21 (1995), pp. 141¯150.

Loftus, 1997. E.F. Loftus , Researchers are showing how suggestion and imagination can create "memories" of events that did not actually occur. Scientific American 277 (1997), pp. 51¯55.

Nijenhuis et al., 1998. E.R.S. Nijenhuis, P. Spinhoven, R. Van Dyck, O. Van der Hart and J. Vanderlinden , Degree of somatoform and psychological dissociation in dissociative disorders is correlated with reported trauma. Journal of Traumatic Stress 11 (1998), pp. 711¯730.

Romkens, 1992. Römkens, R. G. (1992). Gewoon geweld?: Omvang, aard, gevolgen en achtergronden van geweld tegen vrouwen in heteroseksuele relaties. (Just violence?: Scope, nature, consequences and backgrounds of violence against women in heterosexual relationships). Amsterdam/Lisse: Swets & Zeitlinger BV.

UNESCO, 1976. UNESCO (1976). International Standard Classification of Education (ISCED). Paris: UNESCO.

Walling et al., 1994. M.K. Walling, M.W. O'Hara, R.C. Reiter, A.K. Milburn, G. Lilly and S.D. Vincent , Abuse history and chronic pain in women. II. A multivariate analysis of abuse and psychological morbidity. Obstetrics & Gynecology 84 (1994), pp. 200¯206.

Williams, 1995. L.M. Williams , Recovered memories of abuse in women with documented child sexual victimization histories. Journal of Traumatic Stress 8 (1995), pp. 649¯673.

Wyatt & Peters, 1986. G.E. Wyatt and S.D. Peters , Methodological considerations in research on the prevalence of child sexual abuse. Child Abuse & Neglect 10 (1986), pp. 241¯251.

Zanarini et al., 1989. M.C. Zanarini, J.G. Gunderson, M.F. Marino, E.O. Schwartz and F.R. Frankenburg , Childhood experiences of borderline patients. Comprehensive Psychiatry 30 (1989), pp. 18¯25.